Acetylcholinesterase was measured in amniotic fluid from normal chick embryos and embryos with neural tube defects. Neural tube defects were induced in the chick embryos by three procedures, removal of albumen, mechanical disruption of the closed neural tube or injection of tetanus toxin. The concentration of acetylcholinesterase in amniotic fluid from untreated normal embryos changed throughout the period examined (5-14 days incubation) but was stable at 0.5 Ul-1 over the time period 6-11 days. Amniotic fluid taken from treated embryos with neural tube defects at 8 days always contained a higher concentration of acetylcholinesterase than fluid from sham operated but otherwise normal embryos, mean 40.9 Ul-1, S.E.M. = 10.1 Ul-1, versus 1.0 Ul-1, S.E.M. = 0.2 Ul-1. The range of values (6.1-393 Ul-1) was clearly separated from the normal values, range 0.0-5.5 Ul-1. In 13 cases with developmental abnormalities other than neural tube defects, the concentration of acetylcholinesterase was elevated in only one. Two different forms of acetylcholinesterase, as shown by gel electrophoresis, were present in fluid form both normal and defective embryos. These forms were also present in blood plasma, cerebrospinal fluid and in the high speed supernatant from brain extracts, the latter tissue contained an additional form of greater electrophoretic mobility. After irreversible inhibition, enzyme activity in amniotic fluid recovered slowly; only half the control value was reached by 140 h compared with complete recovery in the tissues of the embryo within 19 h. Histochemical staining for acetycholinesterase showed that the spinal cord in the region of the lesion contained high concentrations of the enzyme. The possible sources of acetylcholinesterase in amniotic fluid are discussed. This chicken model of neural tube defects provides support for the use of acetylcholinesterase tests in the detection of neural tube defects clinically, and provides a model for experimentation with this system.